Background: The aim of this study was to describe period and cohort effects in incidence and mortality of stomach and pancreatic cancer in England and Wales.
Methods: National figures for mortality (1951-2000) and incidence (1971-2000) were analysed using log-linear Poisson regression models to obtain relative risks (RR) for period (year of incidence or death) and cohort (year of birth).
Results: Stomach cancer shows a pronounced cohort effect in mortality with a decline in RR in men from 2.20 (1876) to 0.47 (1946) and a reduction from 2.79 to 0.41 for women. Mortality to incidence ratios are now less than 0.70. Pancreatic cancer mortality (men) RR rose from 0.91 (1951-1955) to a peak 1.11 (1976-1980) and then declined to 0.90 (1996-2000). Women showed a similar pattern. Cohort RR (men) increased to a peak of 1.14 in 1916 and declined to 1.01 in 1946, and continued to fall; the peak occurred slightly later in women. Mortality to incidence ratios were near 1 in the first 20 years, declining to 0.95 in the last 10 years.
Conclusion: Stomach cancer incidence has fallen continuously from 19(th) century birth cohorts onwards. Incidence of pancreatic cancer has fallen in successive birth cohorts after 1920; peak period risk was 1976-1990. Age-standardized mortality and case mortality for pancreatic cancer are declining.
Copyright (c) 2007 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.